This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Serious SPEC CPU 2006 FP performance regressions on IA32
- From: "H. J. Lu" <hjl at lucon dot org>
- To: "Menezes, Evandro" <evandro dot menezes at amd dot com>
- Cc: gcc at gcc dot gnu dot org, ubizjak at gmail dot com, Dwarak Rajagopal <dwarak dot rajagopal at txexmtae dot amd dot com>, "Meissner, Michael" <michael dot meissner at amd dot com>
- Date: Mon, 11 Dec 2006 10:09:29 -0800
- Subject: Re: Serious SPEC CPU 2006 FP performance regressions on IA32
- References: <1449F58C868D8D4E9C72945771150BDFDF537C@SAUSEXMB1.amd.com>
On Mon, Dec 11, 2006 at 11:35:27AM -0600, Menezes, Evandro wrote:
> HJ,
>
> > > Gcc 4.3 revision 119497 has very poor SPEC CPU 2006 FP performance
> > > regressions on P4, Pentium M and Core Duo, comparing aganst
> > > gcc 4.2 20060910. With -O2, the typical regressions look like
> > >
> > > Gcc 4.2 Gcc 4.3
> > > 410.bwaves 9.89 9.14 -7.58342%
> > > 416.gamess 7.17 7.16 -0.13947%
> > > 433.milc 7.68 7.65 -0.390625%
> > > 434.zeusmp 5.57 5.55 -0.359066%
> > > 435.gromacs 3.99 4.02 0.75188%
> > > 436.cactusADM 4.59 4.50 -1.96078%
> > > 437.leslie3d 5.78 3.98 -31.1419%
> > > 444.namd 6.25 6.18 -1.12%
> > > 447.dealII 11.3 11.3 0%
> > > 450.soplex 8.61 8.59 -0.232288%
> > > 453.povray 6.70 6.72 0.298507%
> > > 454.calculix 2.81 2.74 -2.4911%
> > > 459.GemsFDTD 6.07 4.95 -18.4514%
> > > 465.tonto 4.44 4.45 0.225225%
> > > 470.lbm 10.6 10.7 0.943396%
> > > 481.wrf 4.56 4.50 -1.31579%
> > > 482.sphinx3 11.2 11.1 -0.892857%
> > > Est. SPECfp_base2006 6.42 6.15 -4.20561%
> > >
> > > Evandro, what do you get on AMD?
> > >
> > > Is that related to recent i386 backend FP changes?
>
> Here's what we got:
>
> ?%
> CPU2006
> 410.bwaves -6%
> 416.gamess
> 433.milc
> 434.zeusmp
> 435.gromacs
> 436.cactusADM
> 437.leslie3d -26%
> 444.namd
> 447.dealII
> 450.soplex
> 453.povray
> 454.calculix
> 459.GemsFDTD -12%
> 465.tonto
> 470.lbm
> 481.wrf
> 482.sphinx3
>
> Though not as pronounced, definitely significant.
>
It is close to what we see on both x86 and x86-64. Are you going to
track it down?
H.J.